Process and Machine Problems
Listed below are the results of five of the most common machine-crash situations:
1. Hot-leak caused 14 days downtime and very high service-specialist cost
2. An RF-etch failure caused total damage to 400 wafers
3. HV spikes (‘glitches’) caused partial damage to 700 wafers
4. Heater temperature-control failure caused damage to hundreds of wafers
5. Operator that selected the wrong process caused the scrap of 500 wafers
As an example, only the damage caused by RF-etch failure was computed:
|Cost of scrapped wafers – 400 wafers @ $300/wafer||$120,000|
|Machine downtime to find the cause and fix – 15 hours @ $500/hour||$7,500|
|Process and maintenance work – 30 hours @ $100/hour||$3,000|
|Cost of test wafers – 25 wafers @ $20/wafer||$500|
|Delayed product, re-catch production, est. – 400 wafers @ $50/wafer||$20,000|
|Other equipment idling, 1% low overall fab utilization 336 hours @ $100/hour||$33,600|
|Total one-event loss||$184,600|
The TadiGuard Solutions
After the catastrophic occurrence, a TadiGuard was immediately installed. The RF bias, as well as the forward and reflected power, are now monitored continuously, avoiding the process when the RF-etch is absent.
In addition, the TadiGuard locks the machine in order to avoid wrong processes. It monitors and keeps the wafer temperature within boundaries, fully tracks the vacuum and its rate-of-change, and inhibits all the above mentioned potential crash situations. All these benefits for the cost of less than ten percent of the loss caused by one event!
- “Crashes” are now minimized.
- Less engineering work, test wafers, tests, and maintenance adjustments are needed.
Over-All Fab Yield, Capacity And Profitability Increased!
Typical Varian 3290 screens
Zoomed in display shows typical pumpdown curve and pumps / valve sequence
Simple 3290 user-drawn real-time machine Mimics screen